Automatic Classification of Skin Cancer Using KNN, SVM and CNN

Abstract

A skin cancer classification system has been designed and developed. This work
presents a new approach to the automated classification of skin cancer images based on
texture and colour features. To remove the unwanted noises in the skin image, median
filtering is used. In the next stage gray level co-occurrence matrix and colour features are
extracted. Finally, k-nearest neighbour, support vector machine and convolutional neural
networks are used to classify the skin cancer images. The application of the proposed
method for tracking skin cancer is demonstrated to help pathologists distinguish its type of
skin cancer. A classification with an accuracy of 85%, 96% and 98% has been obtained by,
k-nearest neighbour support vector machine and convolutional neural networks